Pulmonary emphysema subtypes defined by unsupervised machine learning on CT scans
Elsa Angelini
(1, 2)
,
Jie Yang
(3, 4)
,
Pallavi Balte
(5)
,
Eric Hoffman
(6)
,
Ani Manichaikul
(7)
,
Yifei Sun
(5)
,
Wei Shen
(8, 5)
,
John Austin
(9, 5)
,
Norrina Allen
(10)
,
Eugene Bleecker
(11, 12)
,
Russell Bowler
(13)
,
Michael Cho
(14)
,
Christopher Cooper
(15)
,
David Couper
(16)
,
Mark Dransfield
(17, 17)
,
Christine Kim Garcia
(5)
,
Meilan Han
(18)
,
Nadia Hansel
(19)
,
Emlyn Hughes
(4)
,
David Jacobs
(20)
,
Silva Kasela
(4, 21)
,
Joel Daniel Kaufman
(22)
,
John Shinn Kim
(5, 7)
,
Tuuli Lappalainen
(5)
,
Joao Lima
(19)
,
Daniel Malinsky
(5)
,
Fernando Martinez
(23)
,
Elizabeth Oelsner
(5)
,
Victor Ortega
(24)
,
Robert Paine
(25)
,
Wendy Post
(19)
,
Tess Pottinger
(5)
,
Martin Prince
(23)
,
Stephen Rich
(7)
,
Edwin Silverman
(26)
,
Benjamin Smith
(5)
,
Andrew Swift
(5)
,
Karol Watson
(15)
,
Prescott Woodruff
(27)
,
Andrew Laine
(4)
,
R Graham Barr
(5)
1
IMAGES -
Image, Modélisation, Analyse, GEométrie, Synthèse
2 IDS - Département Images, Données, Signal
3 UCAS - University of Chinese Academy of Sciences [Beijing]
4 Columbia University [New York]
5 CUMC - Columbia University Medical Center
6 University of Iowa [Iowa City]
7 University of Virginia
8 Chongqing University [Chongqing]
9 Health Canada - Santé Canada
10 [Northwestern University Medical School] - Feinberg School of Medicine [Northwestern University, Evanston]
11 University of Aberdeen
12 University of Arizona
13 NJH - National Jewish Health
14 BWH - Brigham & Women’s Hospital [Boston]
15 UCLA - University of California [Los Angeles]
16 University of North Carolina
17 UAB - University of Alabama at Birmingham [ Birmingham]
18 Pulmonary and Critical Care, University of Michigan, Ann Arbor, Michigan
19 JHU - Johns Hopkins University
20 EPSS - Department of Earth, Planetary and Space Sciences [Los Angeles]
21 New York Genome Center [New York]
22 University of Washington [Seattle]
23 Weill Medical College of Cornell University [New York]
24 Mayo Clinic
25 University of Utah
26 Brigham and Women's Hospital [Boston]
27 UC San Francisco - University of California [San Francisco]
2 IDS - Département Images, Données, Signal
3 UCAS - University of Chinese Academy of Sciences [Beijing]
4 Columbia University [New York]
5 CUMC - Columbia University Medical Center
6 University of Iowa [Iowa City]
7 University of Virginia
8 Chongqing University [Chongqing]
9 Health Canada - Santé Canada
10 [Northwestern University Medical School] - Feinberg School of Medicine [Northwestern University, Evanston]
11 University of Aberdeen
12 University of Arizona
13 NJH - National Jewish Health
14 BWH - Brigham & Women’s Hospital [Boston]
15 UCLA - University of California [Los Angeles]
16 University of North Carolina
17 UAB - University of Alabama at Birmingham [ Birmingham]
18 Pulmonary and Critical Care, University of Michigan, Ann Arbor, Michigan
19 JHU - Johns Hopkins University
20 EPSS - Department of Earth, Planetary and Space Sciences [Los Angeles]
21 New York Genome Center [New York]
22 University of Washington [Seattle]
23 Weill Medical College of Cornell University [New York]
24 Mayo Clinic
25 University of Utah
26 Brigham and Women's Hospital [Boston]
27 UC San Francisco - University of California [San Francisco]
Elsa Angelini
- Fonction : Auteur
- PersonId : 1065314
- IdHAL : elsa-angelini
- ORCID : 0000-0002-1602-300X
Ani Manichaikul
- Fonction : Auteur
- PersonId : 1325317
- ORCID : 0000-0002-5998-795X
Michael Cho
- Fonction : Auteur
- PersonId : 801596
- ORCID : 0000-0002-4907-1657
Christine Kim Garcia
- Fonction : Auteur
- PersonId : 1325318
- ORCID : 0000-0002-0771-1249
Joel Daniel Kaufman
- Fonction : Auteur
- PersonId : 1325319
- ORCID : 0000-0003-4174-9037
John Shinn Kim
- Fonction : Auteur
- PersonId : 1325320
- ORCID : 0000-0002-8887-150X
Résumé
Background Treatment and preventative advances for chronic obstructive pulmonary disease (COPD) have been slow due, in part, to limited subphenotypes. We tested if unsupervised machine learning on CT images would discover CT emphysema subtypes with distinct characteristics, prognoses and genetic associations. Methods New CT emphysema subtypes were identified by unsupervised machine learning on only the texture and location of emphysematous regions on CT scans from 2853 participants in the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS), a COPD case–control study, followed by data reduction. Subtypes were compared with symptoms and physiology among 2949 participants in the population-based Multi-Ethnic Study of Atherosclerosis (MESA) Lung Study and with prognosis among 6658 MESA participants. Associations with genome-wide single-nucleotide-polymorphisms were examined. Results The algorithm discovered six reproducible (interlearner intraclass correlation coefficient, 0.91–1.00) CT emphysema subtypes. The most common subtype in SPIROMICS, the combined bronchitis-apical subtype, was associated with chronic bronchitis, accelerated lung function decline, hospitalisations, deaths, incident airflow limitation and a gene variant near DRD1 , which is implicated in mucin hypersecretion (p=1.1 ×10 −8 ). The second, the diffuse subtype was associated with lower weight, respiratory hospitalisations and deaths, and incident airflow limitation. The third was associated with age only. The fourth and fifth visually resembled combined pulmonary fibrosis emphysema and had distinct symptoms, physiology, prognosis and genetic associations. The sixth visually resembled vanishing lung syndrome. Conclusion Large-scale unsupervised machine learning on CT scans defined six reproducible, familiar CT emphysema subtypes that suggest paths to specific diagnosis and personalised therapies in COPD and pre-COPD.